Re: [PATCH v2 3/4] ARM: dts: exynos5250: Fold common ChromeOS parts into Snow

["Andreas Färber" <afaerber@suse.de>]

Hi Javier and Doug,

Am 25.07.2014 19:02, schrieb Javier Martinez Canillas:
> On Fri, Jul 25, 2014 at 6:43 PM, Doug Anderson <dianders@chromium.org> wrote:
>> On Fri, Jul 25, 2014 at 9:35 AM, Javier Martinez Canillas
>> <javier@dowhile0.org> wrote:
>>> Hello Andreas,
>>>
>>> On Fri, Jul 18, 2014 at 7:20 PM, Andreas Färber <afaerber@suse.de> wrote:
>>>>
>>>> +       memory {
>>>> +               reg = <0x40000000 0x80000000>;
>>>> +       };
>>>> +
>>>> +       chosen {
>>>> +       };
>>>> +
>>>
>>> Is there a reason for an empty chosen node? Same for the Spring DTS.
>>
>> I know that the bootloader (U-Boot) fills this in.  I'd have to double
>> check how the bootloader handles the node not being there to begin
>> with.
> 
> Yes, U-Boot fills this but if the node does not exist in the FDT, it
> creates one. Take a look at fdt_chosen() in common/fdt_support.c [0].
> 
> So I think it's better to just remove it since is empty.

Hm, in different context it was stated that device trees shouldn't rely
on bootloader behavior (for a /memory node on Zynq we ended up
specifying the size cell with the real value despite U-Boot overriding
it to a lesser value).

Can we instead settle on filling this node with defaults? :)
bootargs = "console=tty1"; would be my choice for Spring. Would that be
applicable for Snow as well?
Not sure how to specify that via linux,stdout-path property.

I believe this would be a follow-up patch either way.

Since there's at least two series out there trying to fiddle with
-cros-common.dtsi, including one that drops the slot sub-node of the MMC
and one that adds CPU power supply, it would be nice if we can get the
-cros-common parts sorted out and applied. Can you ack/review 1-2?
Should I squash them in a v3?

Thanks,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg